The Contact Centre Podcast: Episode Five
In this episode, Martin Hill-Wilson of Brainfood Consulting discusses how we can improve the customer experience by designing simple, low-effort and fast customer journeys.
As part of our discussion, we also talk about understanding customer behaviours, reducing service costs and the fundamentals of digital transformation.
To listen to the podcast directly from this webpage, just hit the play button below:
This podcast was made possible by our sponsor, Genesys. We now have a new link to visit their website, instead of the link mentioned in the podcast to request a demo.
So, to find out more about Genesys, simply visit their website
Podcast Time Stamps
- 3:17 – The Barriers to Great Customer Journeys
- 7:24 – Changing Customer Habits
- 9:46 – Bad Digital Customer Journey Examples
- 17:31 – Adding Self-Service to Customer Journeys
- 19:15 – Getting Started With Improving Journeys
- 25:48 – Do Chatbots Really Remove Friction?
- 31:47 – Key Takeaways
Here is a Transcript to the Podcast
Jonty Pearce: So what are some of the exciting projects you’ve been working on recently?
Martin Hill-Wilson: My main one is the emotion management for contact centres. It’s a five-workshop program, stretching over six months. I’ve got 35 people on it. We just had a fascinating session last week actually, exploring the history of emotion and the different competing theories and also the psychology behind it. And the people on the course are busily building their own strategy and plan. So that’s a real highlight for me. I’m enjoying that a lot. I’m just doing a little bit of work at the moment looking actually at probably things we are going to be talking about with a big client on how their call centre infrastructure works and journeys and joined-up journeys and all of that good stuff. And I’m starting to think a little bit more actually about the practicalities of trying to go digital. In practice, a lot of people I think are thinking about in the wrong way. So, lots going on. I’m doing my normal speeches and conferences and all of that. So that’s all happening in the background as well. Just about to tack another couple of chapters onto the emotion management series. So that’s all progressing as well. So it’s a busy time. Thank you.
Jonty: Excellent. So this week we’re going to be looking at the topic of providing better customer journeys. So, Martin, what do you think are the biggest barriers to companies in providing better customer journeys?
Martin: Oh, I think the foundation one is that the journeys aren’t good enough, to be rather flippant about it. And the reason is, if there is a difference between a journey and a process, one is orientated to the company’s benefit and the other is orientated to the customer’s. I still think that too many of those journeys are incomplete from the customer’s perspective. The net result is doing things is difficult, effortful and still requires a lot of human intervention.
Jonty: And when you said there’s a process and there’s a journey, the company looks at the journey and the customer looks at the process?
Martin: Yes, I mean, to give an image of that. So we may well build a process around getting something done like onboarding a customer or filling an application. The things that matter to the company around completeness of data, for example, or something of that nature. They’re all perfect from that perspective, but when the customer looks at it, they may well be irritated because that data, for instance, has been provided many times over already and yet you’re still expecting to me to do it again. It may well be, as is often the case now, that journeys are split between physical activity like providing visual evidence of who I am or filling in a document, posting it. The company themselves then has to probably put it through a scanning process and assessment process. But then part of that journey is online and difficult. But the net effect is that the time it takes to get something done is a matter of days, maybe even a matter of weeks as opposed to the fact that if it was all within a digital domain, it could be done within a matter of minutes. And that’s a real effort from the customer’s perspective as well.
Martin: So journeys, generally speaking, are still half-cooked in many respects. They are also complicated because the underlying support that systems provide to complete those things are often siloed in the sense that there may be duplication of effort around gathering the data, inputting the data. It may even be that certain steps need to be repeated in a journey. But if you look at it overall, if you were to design a way of doing something from the word go, simple, low-effort is seldom the way that you would describe lots of the key journeys that contact centres deal with, like onboarding, claims management, fixing something, etc. etc.
Martin: And so the state of customer journeys is, I think, one of the key problems that we’ve got and it matters because traditionally what we’ve done is use humans to help navigate through that complexity. You phone up, you sit in a queue and the person isn’t just helping you. In fact, I can give you a really good example. There was somebody talking to me yesterday about this, saying that when they actually talk to advisors, there was a very simple job which was booking an annual service visit. Now that at face value looks like you can do it entirely through an app, through self-service. It really doesn’t require anything else. But that really wasn’t the reason that the time has been taken or indeed why people are wanting to talk to people. The real problem was to do with the fact that the service engineer had not turned up on time, that the customer maybe had suddenly had an emergency and they needed to change the time, or there was some other logistics around that and they were using the live call in order to get that dealt with.
Martin: And so there was no easy provision built into the alternative way of doing it for those customers to deal with those particular issues. So it dragged them back into that live environment. So generally speaking, people are being used often to navigate these journeys to help customers get through them in a way that’s quite unnecessary. If you look at the fundamental task at the centre of what we’re trying to do, which can often be relatively simple, and if it was better designed, could be done independently of human help.
Jonty: So that’s interesting. So the problems are, it takes too much effort, it’s too complicated and it takes too much time. So what we need to get to is simple, low-effort and fast.
Martin: Yes, indeed. Always at the top of the list. And the other thing I would put at the top of the list is, if we’re talking about this in a context of change of habit, in other words, you used to phone us, would you please do this now in a new way on an app, online or something. That for me represents a change of behaviour, a change of habit. There’s a certain amount of cognitive overhang on that. God, all right then, but if you’re going to make me do that, there really does need to be a reward around that. And the reward basically is that you will, as you’ve just said, make my life somehow better. And that better could be that it’s easier, it’s faster, it’s more convenient as far as my agenda is concerned. However, many of those journeys are being swapped out more for cost-saving rather than the customer-experience benefits.
Martin: The top of the list for me is are we designing something that from the customer’s point of view is abundantly and obviously better? Because if it’s not better, the other problem we’ve got with customer journey management in this sense is that customers won’t change their behaviour and they’re going to stick to the tried and the tested, which still involves sitting in queues and talking to people. And quite frankly, I do not see a great deal of difference, whether that’s delivered via text media or voice media. At the end of the day, you’re still using people inappropriately to do relatively simple things which customers would like to be able to do at their convenience. But it does, as you say, need to be simple, effortless and effective.
Jonty: So, it’s quite interesting there. So the problem with companies wanting to move digital is they’re trying to fundamentally, it sounds like, to reduce cost. And the problem is that the digital processes haven’t been designed rightly, rather than there’s a problem with digital channels sitting in isolation from the contact centre.
Martin: Correct. Absolutely correct. So, it’s very irritating when you discover this point, but unfortunately it’s true. There is a business about using the technology correctly. That’s in a separate part of the discussion. But even if you’ve got great tech, even if you’ve designed it well, the foundation problem is… You see, the interesting thing, particularly digital really means at the end of the day, more often than not, not doing it with humans. That’s the underlying assumption that we’re doing here. And what we have not managed to do is think through in sufficient detail what that human being is in fact providing.
Jonty: A good example of that would be, my son has recently joined one of the new challenger banks, I think it was described as the darling of the Fintech world, one of these companies, and he accidentally ordered two debit cards from them because I think there was a problem with the website and he needed to cancel one of them. So he went on the website and they said, do you want to chat with a chatbot? Sure. And so he typed in, how do I cancel a second card? And the chatbot said, “oh, you can order a second card through the order a second card option on the app. Did this solve your problem?” “No,” he said, “I accidentally ordered a second card and I need to cancel it.” So change the question. “You can order a second card through the order a second card option on the app. Did this solve your problem?” “No.” “Okay, I’ll transfer you to somebody who can help you.” And then it took about five minutes later, he managed to eventually get the problem solved with a person.
Jonty: So I guess the fundamental problem here was there was no cancel a second card option, but neither did the chatbot understand that the word “cancel” overrode “second card” and just looked through a list of frequently asked questions, which is “how do I order a second card?” And it just matched that, which just seemed to make the customer journey an awful lot worse.
Martin: And there was no benefit at all. So that is a good example of the core sets of problems that we’ve got. So in another world, in another way, and in fact what ended up happening is the human being listened, understood and recognized the real issue your son had and then got it fixed, right? So we’re back to square one again, which is that right now, although there was a queue involved and there’s a cost involved in that, nonetheless that outcome was fixed. So now, how could that have been done better? Well, there’s a number of things sitting in there that need to be addressed. The first one is if you’re going to say, right, we really want to try to enable the customer to be autonomous and to self-serve most of the time in terms of managing their financial affairs with us, the first thing is that you’ve got to be extremely clear with customers what it can and it cannot do.
Martin: And I think that in the journey that we’re all going on at the moment, one of the things that brands are absolutely negligent about is broadcasting and promoting and signposting what we can and we cannot do in different ways. In other words, to avoid the frustration and the scenario that you have just described because, indeed, if you are going to turn away from a human being doing something to something like a chatbot, the reality is you are working intent by intent, slowly, in terms of gathering all of the different use cases together, and it’s not going to be the case for a good number of years that you have got a general all-purpose conversational AI that can really answer anything. You’re going to have things prioritized.
Martin: Now what they would have probably prioritized, given their objectives as a business, will be to sell you a new current account or to look to cross-sell those kinds of things, or probably listening to the chat transcripts in the call centre, they would have picked off the top two or three issues that they will most frequently be dealing with through people and they will have tried to have gotten hold of those intents. The different ways in which customers ask about those intents and then provided a chatbot solution. The real problem sitting there is that there’s always going to be the exception to the rule, and it may well be the fact that somebody’s duplicated an order for another card doesn’t happen very often. It might be statistically very infrequent. However, the issue then is how are you going to then make it extremely simple for customers to get through to what they actually require?
Martin: And there are a number of different ways of doing that. There are some solutions, for example, that have still always got live assistance sitting behind it and any indication of frustration from the customer, duplication of questions, immediately causes a barge-in with a live advisor to save all of that problem, and then the customer doesn’t have to sit in a queue and wait to be dealt with as far as that’s concerned. Or it might be that you design it in a certain way that as soon as there is no satisfactory answer, the escalation is immediate and seamless as far as that’s concerned rather than, again, inaccurately identifying the intent and taking people around the loop. But that is not a problem with the technology. It’s more an issue of conversational design. And if we go back a generation to IVR, again, the criticism that many people have got of IVR isn’t actually to do with the technologies, it’s to do it with the design of the technology and often the necessary expertise hasn’t been brought to bear.
Martin: So yesterday, interestingly, I was watching Google’s conversational design team talk about their whole approach to design. And one of the things that comes through very clearly on that is that it’s not… There’s a whole additional stream of activity and thought and preparation that has to be gone through in order to make these things work. And one of the things is that we have a lot of frustration very quickly if we are not understood as customers. Now, human to human, we immediately adapt and mirror that and understand it, either duplicate or mirror it and acknowledge it or probe with an additional question. But that kind of coping strategy that humans do is not being adequately reflected in self-service routines, either as an FAQ or indeed as a conversational AI. So there’s a number of things to get wrong.
Martin: One, as we said, actually at the top of this discussion, the journey itself is complicated, still requiring human intervention. Two, the technology is just not well deployed in terms of its capability to recognize intent and provide answers. And then thirdly, the conversational design component really hasn’t been well done. And some of that, to come back to our key theme, is missing the point around how you think about a journey in its fullness. So I can give you back an example from again, a story I heard two days ago, which was that we used a digital app, this particular person, to get a refund because they needed that. Now interestingly, the refund, though, in terms of the legacy system, could not be completed for five full days. That was not communicated to the customer, nor was there any proactive messaging. The net result, guess what? is that that customer then phones the call centre and says, “where is my money? Did that transaction even work?”
Martin: So, what they failed to do was to really think the fullness of the before, the during and the after, in terms of the key things that the customer will go through. The net effect is they technically have an answer to it, but they have not really fulfilled the needs the customer has got, particularly given the fact that they are causing some of that problem. So customer design, particularly when you’re going digital and you’re going self-service, really has got to take a much broader view of what currently takes place between customer and live human, and you’re trying to mirror that in an entirely self-service way and you have to take a very broad view of all the gotchas and the things that don’t work. Otherwise, you’ve made no progress at all in terms of your twin goals of reducing your service costs and improving the customer experience. All you’ve managed to do is to frustrate people and then put them back into the queue at the end of the day.
Martin: So I think it’s because we’re in very early days on all of this and we really haven’t come to grips with the real complexity of mirroring human behaviour in an autonomous way.
Jonty: So, it looks like a lot of the secret of getting the customer journeys right, as we go forward, is almost how do we get the self-service component designed correctly from the start?
Martin: Yes, I would agree. And some of the truth which is very unfortunate – I was going to say this earlier on – is that your journey is not fit right now to be turned into self-service. It may be a theoretical candidate. You may well have a vendor who says, “I have got an intent library for this particular use case and here’s X, Y, Z who’ve done it.” But your current approach to doing it, if you sit and listen to the way your advisors have to engage with your customers and help them through that journey, it’s a ridiculous idea to imagine a customer is going to be able to do that by themselves. There is too much complexity, it gets people too frustrated, it’s just going to cause people to want to talk to somebody to help them through.
Martin: So therefore a number of journeys, not all of them, but a number of journeys need to be fundamentally reworked from the perspective of simplification and offering a better experience. And that’s generally sitting in that big digital transformation roadmap of trying to make things easier at the end of the day. And once you’ve done that, then actually part and parcel of that redesign is the fact that customers probably want to be able to do that by themselves and are able to do that by themselves. However good the technology is, you can’t tack that on the end of a bad process. We know the old adage about rubbish in, rubbish out on data for CRM, it’s the same point that you can put a great self-service process at the end of a journey and if it’s a bad journey, it’s not going to work.
Jonty: Yeah, I can imagine that’s absolutely key. Where do people get started on improving the customer journey? Is it a pile of Post-it notes and a big process-mapping exercise or is it really just zooming in on one component of service design and designing one enquiry type that people have?
Martin: I think the more granular you get, the better, because funnily enough, it is true, the more you look at these things, that the issues really sit in the detail of it. So I don’t think that you end up with a very large journey map that almost goes from awareness through to support and lifetime value. That’s far too up-in-the-air stuff. You really need to get down to the detail. And I think from a contact centre, customer service perspective, it’s in those things that we do day in, day out, a much more micro-level. So your focus is on the detail of those little things.
Martin: And the question is always to say, “do we really need to be using a human being for that particular task?” I’ve said it many, many times over. I believe that we should only use people going forward when it’s emotional, when it’s complex and when it’s about building the quality of the relationship. And if you just take that as a very simple rule of thumb and then you look over all of your inbound categories, that gives you a very crude way of dividing out what should remain in the human domain and then what should, in theory, be a candidate for another approach.
Martin: Now, there’s a number of things in that. One part of that, by the way, is saying that issue shouldn’t even exist. We’re not going to self-serve that, we’re just going to eliminate it. That’s a silly, redundant way of doing things. That really is failure demand, let’s eliminate. So, that’s one particular conclusion. The next one is, I think, twofold. You either go down the route of robotic process automation whereby the activity really doesn’t warrant human attention quite frankly, it just needs to be automated and that’s either from the perspective of the advisor and or the customer. That’s a whole bunch of stuff and people are beginning to do that. Or actually the customer wants and should be able to do that, but we need to be able to provide them with a sort of logical step-by-step way of doing that. Sometimes it’s easier if it’s designed in a conversational flow and then we’re sort of getting closer towards the world of the bot. But you have to then look at the really detailed stages of, so what’s the first thing the customer needs to do in terms of understanding, in terms of activity, in terms of information they might need to be able to do that activity? So it’s as granular as that.
Martin: And again, the way to understand that task is to, I think, use advisors very, very strongly. Because if they’re currently doing it themselves, what we’re attempting to do is to look at what’s the current way of doing it that we do with a customer? As you do that work with customers, what are the additional considerations that they may be asking you that you are, en passant, also helping to solve? Do you also find that once you’ve done a main task that people say, “oh, and by the way,” and how often does “oh by the way” come up? Which you haven’t anticipated as well as part of your redesign thing. And then look at that and talk about it honestly and say, “the way that we currently do it, is it at all conceivable that we could automate that and allow our customer to do that?”
Martin: And that debate is not too difficult to have. Customer advisors, I think, and people looking at that can say, “no, that’s just not going to happen,” or, “if we modify that, if we simplify that, if we redesigned that or we used what we’ve currently got better.” I mean, one interesting point about self-service is using data that we already have about customers. Again, CRM data is really just not used at all effectively. And so in the redesign process with those criteria that you’ve come up with, which is simple, effortless, the shortest possible route to doing it, maybe some education – that’s the other interesting point, do you need to provide embedded information on how to do stuff? And by the way, is that best written? Is that best done as a little video or something like that? Can we enable customers to do it as far as that’s concerned?
Martin: So I think what you’ve got to do is to take the lowest common denominator, in terms of effort and task completion, and then trial it. The other point about all of this work is that you still don’t really know all the ways the customers can get it right and to get it wrong, so you need a sort of trial test, learn agile, whatever language you want to use, without spending an awful lot of effort doing that stuff. So for me, I think a lot of the redesign happens with paper-based versions, hypothetical things and you just go out and try it with some people. Now again, a lot of people in their own businesses will maybe be customers of that business. That’s a good group of people to start trialling new things on. Trial it with friends and family. Just find out what things do and don’t work there and then you come up with a revised journey with all of those things that the customer is going to find a challenge somehow resolved.
Martin: And that really is the basis of your need that you should have defined before then going to do a self-service solution on top of that. And I think you can do that through a series of very simple workshops with your customer service advisors, team leaders or whatever like that. Think about a better way. Test that out with some customers, come back and then you’ve got the basis of, I think, a much more useful discussion with a vendor who can then help you transform that into a self-service routine.
Jonty: Yes. Because I think one of the things that’s often missed is the testing and the continual improvement of that test before it gets rolled out to everyone.
Martin: Yes. Put the other way around, by the way, is that whilst your son suffered that, what should be unforgivable is that another customer goes through that same experience. So what they really should be doing in the background is having humans watching that. Now, they clearly didn’t manage to intervene in time, but they ought to have recognized that problem and that problem ought to be fixed overnight. Now, there may be a temporary solution there which is to recognize the person who has got a second card by mistake and there is some coping strategy, but very soon after that there ought to be full intent recognition and that that is then seamlessly dealt with. With both, yes, this is how you do it and I will do it for you or we still can’t help you with that, but I will immediately send you through to a person who will fix it for you immediately.
Jonty: I suppose one of the problems with chatbots is that in some ways what you get is people only interact with the chatbots. So you don’t necessarily… People just go, I’m just going to give up. This seems like too much effort. And so, in a way you don’t get the transfer through to the advisor, which might flag up that something’s wrong. It might just be, oh, it was something we didn’t understand.
Martin: I saw a lovely thing last week actually. Chris Ezekiel, who does Creative Virtual, has been doing this for 13-plus years now. He had a lovely little picture on the screen showing the transfer from a virtual assistant, which was in a simple box, just a simple visual avatar of that particular virtual assistant, and then still staying within that same dialogue box was then a picture of a real human being still done in the same kind of visual drawing style. But it demonstrated to the customer that they were moving from a virtual assistant to a live assistant. He said that that really helped the customer recognize that change in the journey. And of course, the trick of it is that everything that the customer will have done with the chatbot to date, that interaction history, will be visible immediately, obviously, for the live agent to understand, pick up and then seamlessly pick it up from that particular point.
Martin: And I think at the end of the day, and we’re not there at all, first competency is to get that seamless escalation to live. And here’s a point, I think that if you have got confidence that you have redesigned a journey which is better than the old, what you should find is that most customers just don’t want to talk to a person if you’ve done a better job and it’s self-service. So therefore, with that confidence, you should always make it visible and easy for a customer to escalate to a live assistant if they require it. And you should find that that is happening very seldom because you’ve designed a great process, but you can never anticipate situations and you shouldn’t try to say customers will always be self-sufficient. So getting access to a live human being should always be made very, very simple.
Jonty: And in many ways that access to a live human being you see most at the supermarket, where their processes for self-service supermarkets seem pretty broken to me. They need some designing. But actually, at most supermarkets you can pretty quickly get an access to… There’s a live person looking after, in some cases, four people, in some cases it’s two people, to help use the self-service system. So they haven’t hidden people and say, “put your hand up, have a five-minute queue and we’ll get someone to… Or just go to another checkout.” Actually have people hovering there waiting to help customers. And it strikes me that in the contact centre, if we are going to automate customer service, we should have people hovering rather… So we should really overstaff that escalation, so we can do it quickly rather than just going, “hey, that’s another queue, let’s do a 80/20 80/60 and wait 60 seconds.”
Martin: Well said, because look, let’s look at the psychology, right? So I’m a call centre director, manager, leader or something like that, and I’ve got no process X journey X redesigned in the way that we’ve just been talking about. So the very first launch should be, as you say, total overstaffing. Well, the very first thing we do is actually let customers know about that, proactively. We then let them know if they’re still coming into live assistance, that maybe here’s another way and let me hold your hand and do it with you. But then as they do it and they build confidence and trust in a new way of doing it, you completely overstaff it, as you have said, and you keep watching it. And every time you overstaff it and you intervene, something has not worked.
Martin: So therefore, you take that learning and you re-engineer something, it’s in the micro-detail of stuff, until that part of that journey continues to have higher and higher levels of effectiveness. And you keep going, keep going, keep going. You still keep people attached to that journey, but once it becomes a mature self-service journey, you’ve maybe got only 10%, 12%, 4%, nor 0.3% of people that need to escalate because the journey has been so well designed. But you could never have anticipated all those little gotchas unless you had got into that continuous human-in-the-loop approach to continuous redesign and smoothing out the journey. That’s how the designers of this world do that stuff. It’s the law of that incremental improvement, marginal gains, it’s all about that stuff. And unless you work it in that way, as you’ve just said, I mean, it’s very instructive actually, what you have said, because if you look to the reasons why that person’s sitting there, there’s the obvious, the one like the alcohol thing: You’ve bought something like a knife or alcohol, you need to have permission to buy that. So, that’s one reason.
Martin: But the number of times that the barcode doesn’t quite work, the number of times I inadvertently have actually duplicated my buying bread more than once, I need that person. If you look at what goes on, there is constant intervention from that person because the customer didn’t quite get that right, and they’ve got it entirely correct by having, you know, zero effort to go and find assistance. A person is always hovering around. So why we imagine, to your point, that self-service and contact centres should really be any different I can’t imagine. So, I don’t think that we’re thinking about how you initiate a new self-service routine, how you understand the inevitable rough edges, how you then smooth those out over time and how you go through essentially the maturing process, and as a leadership, recognize that it’s reached a point of pretty strong optimization and it’s still connected in to live, and you’re still interested in the occasions when people need to use a live assistant.
Jonty: Indeed. So I think some very sound advice there. So, just to summarize a few of the things we’ve talked about today, if we want to improve the customer journey, we need to make it simple, low-effort and fast. And it’s really about the digital transformation of how can we design this process so the customer could take care of it all themselves. That’s all about customer design, which I don’t think we talk about anywhere near enough, and ultimately the devil is in the detail. It is something that companies often don’t like to go through because there’s no, if you like, quick fix in the detail, it’s really drilling down with that micro-approach. So is there a type of person we need to get on to the customer transformation? Somebody who’s very detail orientated, someone very focused on things not being right?
Martin: Yes. I think there are some formal skills like service design we could talk about, and those sorts of disciplines, which I really do think are good to bring in. But in terms of what type of person you’re looking for, it is someone with a bit of an obsession for the detail, for perfection, for being in service to customers, who can really think logically, who’s data literate. Because a lot of this is to do with understanding feedback and getting down to root cause stuff and test and learn and being very systematic about that. I think you have to have stamina and a real desire to get the stuff right. And often, the underlying stuff, of course, is going to be a combination as ever of people, process, tech. So, there’s going to be a certain skill also in getting those owners of those different things to again change behaviour and help make those journeys better.
Martin: But underneath that, there is certainly a core skill, which is not to be satisfied until it works for the customer. So, I’m not so keen on people who are good, for example, at lean and that sort of thing, not because I think they aren’t competent, but they have been orientated to think about the internal agenda taking cost out. And it’s all too easy with these programs to be satisfied that it’s in, it does a good enough job and off we go to the next thing. The way we’ve just been discussing it is much more obsessive and not to be satisfied until it has worked better than the previous way of doing it. So, it’s stubborn customer-focused, detail-orientated persistence. And if you’ve got that, then I think you qualify for taking those kind of digital journeys on their maturity cycle.
Jonty: Absolutely fabulous. So I think that should be a poster that we should put up in our contact centre. Don’t be satisfied until it is good enough for the customer. We should make those up as motivational posters and circulate them around the customer service world. Martin Hill-Wilson, thank you very much for joining us today.
Martin: My pleasure, Jonty. Thank you for inviting me. It’s been a good discussion. Thank you.
Jonty: And that’s all for this episode. Thank you to Martin Hill-Wilson from Brainfood Consulting for joining us today.